pandas布尔索引仍保留原始MultiIndex

Question

我有一个数据框

import numpy as np
import pandas as pd
arrays = [np.array(['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux']),np.array(['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two'])]
df = pd.DataFrame(np.random.randn(8, 4), index=arrays)

df.columns = ['a','b','c','d']

df
Out[3]: 
                a         b         c         d
bar one  0.346640 -0.908057  1.327248 -0.600094
    two -0.623039 -0.146015 -0.295474  0.283444
baz one -0.468552  0.319582  0.293260 -0.329329
    two -0.912441  0.719779 -2.136825  0.997689
foo one -0.839984 -1.186596 -0.458738  0.661190
    two -0.480537  0.514584 -0.284970 -1.871232
qux one  0.079585 -1.062287  0.075252  0.041869
    two -2.285919 -0.697770  0.443770  0.072648

filtered_df = df.loc[
    (df['b'] < -1)
]

当我尝试一些布尔索引时，它可以工作：

filtered_df
Out[5]: 
                a         b         c         d
foo one -0.839984 -1.186596 -0.458738  0.661190
qux one  0.079585 -1.062287  0.075252  0.041869

那么为什么index属性仍然包含第一个DataFrame的所有索引？为什么baz例如？

filtered_df.index
Out[6]: 
MultiIndex(levels=[['bar', 'baz', 'foo', 'qux'], ['one', 'two']],
           labels=[[2, 3], [0, 0]])